From: Demian K. <dem...@vi...> - 2011-03-31 12:46:34
|
Those anomalies are likely related to records with multiple call numbers -- since the Browse feature is built on faceting, it just says "give me all the tens-level call numbers from records that also contain this hundreds-level call number." If everything is single-valued, that works really well, but when you have different call numbers on a single record, you'll see some anomalies. If that doesn't turn out to be the cause of the problem, please let me know and I can investigate further! - Demian > -----Original Message----- > From: Matthew Hooper [mailto:mat...@fl...] > Sent: Wednesday, March 30, 2011 10:21 PM > To: vuf...@li... > Subject: Re: [VuFind-Tech] Dewey with 1.1 > > Hi Demian, > > Call number browse appears to be working fine now after a reindex and > stop/start of the services. There does seem to still be some anomalies > when you drill down to a series like the 800's with some 7xx and 9xx > call numbers appearing in the list, but I can't remember whether that > was happening before the upgrade or not. > > Thanks, > > Matt. > > -- > Matthew Hooper > Systems Officer, > Flinders University Library > G.P.O. Box 2100 > ADELAIDE, South Australia 5001 > Phone: +618 8201 2068 > Fax: +618 8201 2508 > Email: Mat...@fl... > > > -----Original Message----- > From: Demian Katz [mailto:dem...@vi...] > Sent: Wednesday, 30 March 2011 9:39 PM > To: Matthew Hooper; vuf...@li... > Subject: RE: [VuFind-Tech] Dewey with 1.1 > > The fixed SolrMarc was released this morning (thanks, Bob!). You can > download it here: > > http://code.google.com/p/solrmarc/downloads/list > > Just grab the GenericVuFind version and copy the .jar files from the > archive into your vufind/import directory. > > Please let me know if you continue to have problems after this! > > - Demian > ________________________________________ > From: Matthew Hooper [mat...@fl...] > Sent: Tuesday, March 29, 2011 9:19 PM > To: vuf...@li... > Subject: Re: [VuFind-Tech] Dewey with 1.1 > > Thanks Demian, > > I’ve turned off dewey call number browse for the moment since that’s > where the issue is more apparent, but if you could let the list know > once the solr update has been committed then I’ll grab a fresh copy and > give it a whirl. > > Cheers, > > Matt. > > -- > Matthew Hooper > Systems Officer, > Flinders University Library > G.P.O. Box 2100 > ADELAIDE, South Australia 5001 > Phone: +618 8201 2068 > Fax: +618 8201 2508 > Email: > Mat...@fl...<mailto:Mat...@fl...> > > From: Demian Katz [mailto:dem...@vi...] > Sent: Tuesday, 29 March 2011 11:46 PM > To: Matthew Hooper; vuf...@li... > Subject: RE: [VuFind-Tech] Dewey with 1.1 > > It sounds like there should be a new SolrMarc release in the near > future to fix this problem; if you can wait a week or two, the easiest > solution may be to wait for that and then simply replace the SolrMarc > .jar files with the new version (in fact, if you’re impatient, you > could build it from source now to achieve the same effect – it looks > like the fix was committed to the SolrMarc trunk yesterday). > > If you don’t want to deal with SolrMarc right now, the other option is > to uncomment the dewey.bsh-related lines in > import/marc_local.properties and edit the BeanShell code in > import/index_scripts/dewey.bsh to change all the “if > (CallNumUtils.isValidDewey(current))” lines to “if (true)”. Of course, > if you have bad data in the Dewey fields, this may let it through and > cause a different set of problems… but it’s the quickest patch that > doesn’t involve upgrading SolrMarc. > > Apologies for the inconvenience, and please let me know if you still > have questions. > > - Demian > > From: Matthew Hooper [mailto:mat...@fl...] > Sent: Tuesday, March 29, 2011 3:05 AM > To: vuf...@li... > Subject: Re: [VuFind-Tech] Dewey with 1.1 > > Hi Demian, > > I’ve just run up 1.1 with Dewey support, and can confirm that there’s > something going on with how the records are imported. Our call numbers > are stored in the $984c), and if you take a look at the call number > browse screen at http://library.flinders.edu.au/vufind/ you can see > there are large chunks missing from the indexed call number ranges. Did > you have those instructions handy on how to change the dewey parsing > routines to be more lenient? That way I can do a reload and see if it > brings back those records. > > Thanks, > > Matt. > > -- > Matthew Hooper > Systems Officer, > Flinders University Library > G.P.O. Box 2100 > ADELAIDE, South Australia 5001 > Phone: +618 8201 2068 > Fax: +618 8201 2508 > Email: > Mat...@fl...<mailto:Mat...@fl...> > > From: Demian Katz [mailto:dem...@vi...] > Sent: Wednesday, 23 March 2011 6:30 AM > To: ho...@hs... > Cc: vuf...@li... > Subject: Re: [VuFind-Tech] Dewey with 1.1 > > Okay, I see the problem now – SolrMarc includes a Dewey validation > function which checks call numbers against a regular expression before > indexing them. It expects at the very minimum, three digits followed > by a letter. Your MARC data only has three digits – no letter. > > Two possible solutions: > > > 1.) Is there data missing from your MARC export? Can you get more > detailed Dewey numbers? > > 2.) If you can’t change the data, you can revert to using the > dewey.bsh script in place of the built-in SolrMarc custom function > (restoring the lines I just told you to remove from > marc_local.properties), and you can change the > CallNumUtils.isValidDewey() calls in dewey.bsh to something else that > allows more forgiving validation of the Dewey numbers. > > It appears that the Dewey validation logic was made more strict with > SolrMarc release 2.2, which explains why the problem has just cropped > up for you. I’ll send a note to the solrmarc-tech list to see what the > reasoning was behind this change, and perhaps the next version of > SolrMarc can restore the more lenient analysis. > > - Demian > > From: John C. Houser [mailto:ho...@pa...] > Sent: Tuesday, March 22, 2011 3:36 PM > To: Demian Katz > Cc: ho...@hs...; vuf...@li... > Subject: Re: [VuFind-Tech] Dewey with 1.1 > > Demian, > We'll definitely close the port, when we're live. This is just a test > server. > > No, I've not modified the marc.properties at all. > > I've attached a sample file with 10 USMARC records in it that have > Dewey numbers in the 082a. I extracted them from the larger test file I > loaded. > > J > ________________________________ > > [cid:image001.jpg@01CBEECD.444F86C0] > > Demian Katz<mailto:dem...@vi...> > March 22, 2011 3:14 PM > > > You will want to firewall off that port once you go live; mischievous > users can do some damage if they can access Solr directly. In the > meantime, though, it’s definitely convenient to be able to share the > link! > > It definitely looks like there is no data getting indexed into the > dewey-tens field – that’s probably a good thing, since it at least > narrows the problem to the import process rather than something more > complicated buried in the guts of VuFind itself. Is your > import/marc.properties file customized at all, or are you using the one > that came with 1.1? Would it be possible for you to share a sample > MARC record containing Dewey data? Perhaps as a sanity check it would > be useful for me to try indexing it here to be sure it works on my end. > > - Demian > > ----------------------------------------------------------------------- > ------- > Enable your software for Intel(R) Active Management Technology to meet > the > growing manageability and security demands of your customers. > Businesses > are taking advantage of Intel(R) vPro (TM) technology - will your > software > be a part of the solution? Download the Intel(R) Manageability Checker > today! http://p.sf.net/sfu/intel-dev2devmar > _______________________________________________ > Vufind-tech mailing list > Vuf...@li...<mailto:Vufind- > te...@li...> > https://lists.sourceforge.net/lists/listinfo/vufind-tech > ________________________________ > > [cid:image001.jpg@01CBEECD.444F86C0] > > John C. Houser<mailto:ho...@pa...> > March 22, 2011 3:07 PM > > > Demian, > > This URL worked: http://vufind101.accesspa.org:8080/solr/biblio/admin/ > > It's open to the public, so you can try it yourself. > > The query you suggested, returned this: > − > <response> > − > <lst name="responseHeader"> > <int name="status">0</int> > <int name="QTime">1</int> > − > <lst name="params"> > <str name="indent">on</str> > <str name="start">0</str> > <str name="q">dewey-tens:[* TO *]</str> > <str name="version">2.2</str> > <str name="rows">10</str> > </lst> > </lst> > <result name="response" numFound="0" start="0"/> > </response> > ----------------------------------------------------------------------- > ------- > Enable your software for Intel(R) Active Management Technology to meet > the > growing manageability and security demands of your customers. > Businesses > are taking advantage of Intel(R) vPro (TM) technology - will your > software > be a part of the solution? Download the Intel(R) Manageability Checker > today! http://p.sf.net/sfu/intel-dev2devmar > _______________________________________________ > Vufind-tech mailing list > Vuf...@li...<mailto:Vufind- > te...@li...> > https://lists.sourceforge.net/lists/listinfo/vufind-tech > ________________________________ > > [cid:image001.jpg@01CBEECD.444F86C0] > > Demian Katz<mailto:dem...@vi...> > March 22, 2011 2:30 PM > > > I doubt this is the cause of your problem, but you can actually remove > all of those dewey lines from marc_local.properties. The Dewey > functionality is now part of the SolrMarc core, so it is no longer > necessary to use the external dewey.bsh file. The default Dewey > behavior in marc.properties does the same thing as your BeanShell-based > configuration in marc_local.properties… only it should be somewhat > faster. > > I also recommend getting rid of all the strip_trailing_dot stuff at the > bottom of the file – the punctuation cleanup is now handled internally > by Solr, so you don’t need to do it in SolrMarc. Also, the era field > has been renamed to era_facet, so that may be causing unexpected > behavior. Again, probably not the cause of your current specific > problem, but worth cleaning up and reindexing. > > Is it possible that a problem was introduced when you customized the > Solr schema? I see that you have a custom series_facet field. Did you > manually re-add that in 1.1? > > Have you tried looking at the index in the Solr admin panel (Probably > available at http://your_vufind_server:8080/solr/biblio)? If you do a > search for > > dewey-tens:[* TO *] > > …what do you get? > > - Demian > > ----------------------------------------------------------------------- > ------- > Enable your software for Intel(R) Active Management Technology to meet > the > growing manageability and security demands of your customers. > Businesses > are taking advantage of Intel(R) vPro (TM) technology - will your > software > be a part of the solution? Download the Intel(R) Manageability Checker > today! http://p.sf.net/sfu/intel-dev2devmar > _______________________________________________ > Vufind-tech mailing list > Vuf...@li...<mailto:Vufind- > te...@li...> > https://lists.sourceforge.net/lists/listinfo/vufind-tech > ________________________________ > > [cid:image001.jpg@01CBEECD.444F86C0] > > John C. Houser<mailto:ho...@pa...> > March 22, 2011 2:21 PM > > > Demian, > I migrated the import rules. I also tried using the default > marc.properties settings. The config.ini has dewey = "1" in the > [Browse] section. The facets.ini settings appear to have migrated okay. > I've pasted in my marc_local.properties below. What other details would > help? > > J > > *** > > # Uncomment the following settings to insert appropriate values for > your site: > collection = "HSLC Evergreen Group" > institution = 852a, evergreen_institution.properties(institution) > building = 852b, evergreen_location.properties(location) > > # Uncomment the following settings to use the .bsh scripts in > import/scripts/ > # instead of the built-in SolrMarc functionality found in the .jar > file. > # (by default, the scripts have the same behavior as the built-in > functions, > # but the external scripts are easier to customize to your needs). > #format = script(format.bsh), getFormat, format_map.properties > #callnumber = script(callnumber.bsh), getFullCallNumber > #callnumber-subject = script(callnumber.bsh), getCallNumberSubject, > callnumber_subject_map.properties > #callnumber-subject-code = script(callnumber.bsh), getCallNumberSubject > #callnumber-label = script(callnumber.bsh), getCallNumberLabel > > # Uncomment the following settings to index Dewey call numbers: > dewey-hundreds = script(dewey.bsh), getDeweyNumber(082a:083a, 100), > ddc22_map.properties(hundreds) > dewey-tens = script(dewey.bsh), getDeweyNumber(082a:083a, 10), > ddc22_map.properties(tens) > dewey-ones = script(dewey.bsh), getDeweyNumber(082a:083a, 1), > ddc22_map.properties(ones) > dewey-full = script(dewey.bsh), getDeweySearchable(082a:083a) > dewey-sort = script(dewey.bsh), getDeweySortable(082a:083a) > > # HSLC Evergreen Series facet > series_facet = custom, removeTrailingPunct(490a) > > # Strip dots only > topic_facet = 600x:610x:611x:630x:648x:650a:650x:651x:655x, > (pattern_map.strip_trailing_dot) > genre_facet = 600v:610v:611v:630v:648v:650v:651v:655a:655v, > (pattern_map.strip_trailing_dot) > geographic_facet = 600z:610z:611z:630z:648z:650z:651a:651z:655z, > (pattern_map.strip_trailing_dot) > era = 600d:610y:611y:630y:648a:648y:650y:651y:655y, > (pattern_map.strip_trailing_dot) > pattern_map.strip_trailing_dot.pattern_0 = (.*)\\.=>$1 > pattern_map.strip_trailing_dot.pattern_1 = (.*[^\\.]*)=>$1 > ----------------------------------------------------------------------- > ------- > Enable your software for Intel(R) Active Management Technology to meet > the > growing manageability and security demands of your customers. > Businesses > are taking advantage of Intel(R) vPro (TM) technology - will your > software > be a part of the solution? Download the Intel(R) Manageability Checker > today! http://p.sf.net/sfu/intel-dev2devmar > _______________________________________________ > Vufind-tech mailing list > Vuf...@li...<mailto:Vufind- > te...@li...> > https://lists.sourceforge.net/lists/listinfo/vufind-tech > ________________________________ > > [cid:image001.jpg@01CBEECD.444F86C0] > > Demian Katz<mailto:dem...@vi...> > March 22, 2011 2:05 PM > > > Did you migrate your SolrMarc import rules? If you customized > import/marc.properties or import/marc_local.properties in your RC2 > installation, you may have to copy those changes into the new > installation. By default, VuFind 1.1 should be populating the Dewey > fields using 082a and 083a; if you rely on any other MARC fields, that > might explain your problem. > > Also, did you double check that the settings in facets.ini, > searches.ini and config.ini migrated over correctly? Dewey options are > off by default in 1.1; if you ran the upgrade script, that should have > brought your old settings over, but if something went wrong, it’s > possible that some configurations defaulted back to being disabled. > > Let me know if you still need help after checking these things – I did > check Dewey support before making the release, so I’m pretty sure > nothing is fundamentally broken… but any additional details would be > appreciated if none of these ideas turn out to provide a simple fix! > > - Demian > > From: John C. Houser [mailto:ho...@pa...] > Sent: Tuesday, March 22, 2011 1:22 PM > To: vuf...@li...<mailto:vufind- > te...@li...> > Subject: [VuFind-Tech] Dewey with 1.1 > > Guys, > > I just installed VuFind 1.1, which I migrated from an older RC2 > installation. Now that I've migrated our theme and the few very minor > code changes we've made, everything works except the Dewey call number- > related stuff. We were using the dewey-tens index for a facet and also > as a limit in the advanced search and in the browsing. The dewey index > has disappeared from everything. > > Any ideas what I should be checking to fix this? > > J > -- > John Houser > Technology Coordinator > HSLC > 215-534-6820 > ho...@hs...<mailto:ho...@hs...> > > ----------------------------------------------------------------------- > ------- > Enable your software for Intel(R) Active Management Technology to meet > the > growing manageability and security demands of your customers. > Businesses > are taking advantage of Intel(R) vPro (TM) technology - will your > software > be a part of the solution? Download the Intel(R) Manageability Checker > today! http://p.sf.net/sfu/intel-dev2devmar > _______________________________________________ > Vufind-tech mailing list > Vuf...@li...<mailto:Vufind- > te...@li...> > https://lists.sourceforge.net/lists/listinfo/vufind-tech > ----------------------------------------------------------------------- > ------- > Create and publish websites with WebMatrix > Use the most popular FREE web apps or write code yourself; > WebMatrix provides all the features you need to develop and > publish your website. http://p.sf.net/sfu/ms-webmatrix-sf > _______________________________________________ > Vufind-tech mailing list > Vuf...@li... > https://lists.sourceforge.net/lists/listinfo/vufind-tech |